Re: Database design, Keys and some other things

Let the data from the relation Car be represented with the following relations:
(this is the relation Car from the beginning of the thread)

S-Relation:                    CarKey     CarID
                              -------------------
                                23         vin1
                                24         vin1
                                25         vin1
                                26         vin2
                                27         vin3


K1-Relation:                   CarKey      Make

                             ------------------
                                23        Buick
                                24        Buick
                                25        Buick
                                26        Honda
                                27        Ford


K2-Relation:                  CarKey      Color
                             -------------------
                                23         Silver
                                24         Blue
                                25         Red
                                26         Silver
                                27         Black

Now we can set the query on above S-Relation, K1-Relation and K2-relation and we can get for example the following tuple:

          ( 24, vin1, Buick, Blue )

This is what you want: that part of the tuple identifies tuple. Key "24" uniquely identifies the above tuple which it is part. First key "24" was set in the S-Relation as well as in the K1-Relation and K2-Relation which are real and in the database.
( Similar ideas exist in the math and the computer science)
This Key also has the semantic line, which beginning is one event in the Real World. (see the definition of the key).
(We also can notice that this is not a surrogate key by definition.)

In www.dbdesign10.comm 2.4 under 2. I defined K-Relations, E-relation and
S-Relation.
Here P,E,A1,...,An are the attributes of a relation from RM.
(They are also the pictures of some attributes P, E, A1,...An from the
Conceptual Model)

Here in RM, S-Relation:

a) Determines the keys in the database
b) Shows relationship to entity's identifier E
c) Associates knowledge related to P

I also wrote in my text that we create keys using Constructors that is we can for example click on a button, New or Close to get the Constructor's screen with many fields on it. So I can create a key and the knowledge related to this key: who initialize this Constructor, name of the Constructor (it can be more then one), which procedure created key, what is the previous key,... Now besides above mentioned the database "can recognize" which procedures (logic) created its data. So we can get some meaning and knowledge about the key.If we really need this.
(My notice to dawn about my mistake was put because, the relation was
in form 2.4 under 1)

Vladimir Odrljin Received on Fri Sep 30 2005 - 10:48:37 CEST